summaryrefslogtreecommitdiff
path: root/source3/smbd/oplock_linux.c
blob: 8c4259ac5359700831eaa98bb2efea84211baf04 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
/* 
   Unix SMB/Netbios implementation.
   Version 3.0
   kernel oplock processing for Linux
   Copyright (C) Andrew Tridgell 2000
   
   This program is free software; you can redistribute it and/or modify
   it under the terms of the GNU General Public License as published by
   the Free Software Foundation; either version 2 of the License, or
   (at your option) any later version.
   
   This program is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
   GNU General Public License for more details.
   
   You should have received a copy of the GNU General Public License
   along with this program; if not, write to the Free Software
   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
*/

#include "includes.h"

#if HAVE_KERNEL_OPLOCKS_LINUX

static VOLATILE sig_atomic_t signals_received;
static VOLATILE sig_atomic_t signals_processed;
static VOLATILE sig_atomic_t fd_pending; /* the fd of the current pending signal */

#ifndef F_SETLEASE
#define F_SETLEASE	1024
#endif

#ifndef F_GETLEASE
#define F_GETLEASE	1025
#endif

#ifndef CAP_LEASE
#define CAP_LEASE 28
#endif

#ifndef RT_SIGNAL_LEASE
#define RT_SIGNAL_LEASE 33
#endif

#ifndef F_SETSIG
#define F_SETSIG 10
#endif

/****************************************************************************
handle a LEASE signal, incrementing the signals_received and blocking the signal
****************************************************************************/
static void signal_handler(int sig, siginfo_t *info, void *unused)
{
	BlockSignals(True, sig);
	fd_pending = (sig_atomic_t)info->si_fd;
	signals_received++;
	sys_select_signal();
}

/****************************************************************************
try to gain a linux capability
****************************************************************************/
static void set_capability(unsigned capability)
{
#ifndef _LINUX_CAPABILITY_VERSION
#define _LINUX_CAPABILITY_VERSION 0x19980330
#endif
	/* these can be removed when they are in glibc headers */
	struct  {
		uint32 version;
		int pid;
	} header;
	struct {
		uint32 effective;
		uint32 permitted;
		uint32 inheritable;
	} data;

	header.version = _LINUX_CAPABILITY_VERSION;
	header.pid = 0;

	if (capget(&header, &data) == -1) {
		DEBUG(3,("Unable to get kernel capabilities (%s)\n", strerror(errno)));
		return;
	}

	data.effective |= (1<<capability);

	if (capset(&header, &data) == -1) {
		DEBUG(3,("Unable to set %d capability (%s)\n", 
			 capability, strerror(errno)));
	}
}


/****************************************************************************
call SETLEASE. If we get EACCES then we try setting up the right capability and
try again
****************************************************************************/
static int linux_setlease(int fd, int leasetype)
{
	int ret;

	if (fcntl(fd, F_SETSIG, RT_SIGNAL_LEASE) == -1) {
		DEBUG(3,("Failed to set signal handler for kernel lease\n"));
		return -1;
	}

	ret = fcntl(fd, F_SETLEASE, leasetype);
	if (ret == -1 && errno == EACCES) {
		set_capability(CAP_LEASE);
		ret = fcntl(fd, F_SETLEASE, leasetype);
	}

	return ret;
}


/****************************************************************************
 * Deal with the Linux kernel <--> smbd
 * oplock break protocol.
****************************************************************************/
static BOOL linux_oplock_receive_message(fd_set *fds, char *buffer, int buffer_len)
{
	SMB_DEV_T dev;
	SMB_INO_T inode;
	SMB_STRUCT_STAT sbuf;
	BOOL ret = True;

	if (signals_received == signals_processed) return False;

	if (sys_fstat((int)fd_pending,&sbuf) == -1) {
		DEBUG(0,("Invalid file descriptor %d in kernel oplock break!\n", (int)fd_pending));
		ret = False;
		goto out;
	}

	dev = sbuf.st_dev;
	inode = sbuf.st_ino;
     
	DEBUG(3,("receive_local_message: kernel oplock break request received for \
dev = %x, inode = %.0f\n", (unsigned int)dev, (double)inode ));
     
	/*
	 * Create a kernel oplock break message.
	 */
     
	/* Setup the message header */
	SIVAL(buffer,OPBRK_CMD_LEN_OFFSET,KERNEL_OPLOCK_BREAK_MSG_LEN);
	SSVAL(buffer,OPBRK_CMD_PORT_OFFSET,0);
     
	buffer += OPBRK_CMD_HEADER_LEN;
     
	SSVAL(buffer,OPBRK_MESSAGE_CMD_OFFSET,KERNEL_OPLOCK_BREAK_CMD);
     
	memcpy(buffer + KERNEL_OPLOCK_BREAK_DEV_OFFSET, (char *)&dev, sizeof(dev));
	memcpy(buffer + KERNEL_OPLOCK_BREAK_INODE_OFFSET, (char *)&inode, sizeof(inode));	

 out:
	/* now we can receive more signals */
	fd_pending = (sig_atomic_t)-1;
	signals_processed++;
	BlockSignals(False, RT_SIGNAL_LEASE);
     
	return ret;
}


/****************************************************************************
 Attempt to set an kernel oplock on a file.
****************************************************************************/
static BOOL linux_set_kernel_oplock(files_struct *fsp, int oplock_type)
{
	if (linux_setlease(fsp->fd, F_WRLCK) == -1) {
		DEBUG(3,("set_file_oplock: Refused oplock on file %s, fd = %d, dev = %x, \
inode = %.0f. (%s)\n",
			 fsp->fsp_name, fsp->fd, 
			 (unsigned int)fsp->dev, (double)fsp->inode, strerror(errno)));
		return False;
	}
	
	DEBUG(3,("set_file_oplock: got kernel oplock on file %s, dev = %x, inode = %.0f\n",
		  fsp->fsp_name, (unsigned int)fsp->dev, (double)fsp->inode));

	return True;
}


/****************************************************************************
 Release a kernel oplock on a file.
****************************************************************************/
static void linux_release_kernel_oplock(files_struct *fsp)
{
	if (DEBUGLVL(10)) {
		/*
		 * Check and print out the current kernel
		 * oplock state of this file.
		 */
		int state = fcntl(fsp->fd, F_GETLEASE, 0);
		dbgtext("release_kernel_oplock: file %s, dev = %x, inode = %.0f has kernel \
oplock state of %x.\n", fsp->fsp_name, (unsigned int)fsp->dev,
                        (double)fsp->inode, state );
	}

	/*
	 * Remove the kernel oplock on this file.
	 */
	if (linux_setlease(fsp->fd, F_UNLCK) == -1) {
		if (DEBUGLVL(0)) {
			dbgtext("release_kernel_oplock: Error when removing kernel oplock on file " );
			dbgtext("%s, dev = %x, inode = %.0f. Error was %s\n",
				fsp->fsp_name, (unsigned int)fsp->dev, 
				(double)fsp->inode, strerror(errno) );
		}
	}
}


/****************************************************************************
parse a kernel oplock message
****************************************************************************/
static BOOL linux_kernel_oplock_parse(char *msg_start, int msg_len, SMB_INO_T *inode, SMB_DEV_T *dev)
{
	/* Ensure that the msg length is correct. */
	if (msg_len != KERNEL_OPLOCK_BREAK_MSG_LEN) {
		DEBUG(0,("incorrect length for KERNEL_OPLOCK_BREAK_CMD (was %d, should be %d).\n", 
			 msg_len, KERNEL_OPLOCK_BREAK_MSG_LEN));
		return False;
	}

        memcpy((char *)inode, msg_start+KERNEL_OPLOCK_BREAK_INODE_OFFSET, sizeof(*inode));
        memcpy((char *)dev, msg_start+KERNEL_OPLOCK_BREAK_DEV_OFFSET, sizeof(*dev));

        DEBUG(3,("kernel oplock break request for file dev = %x, inode = %.0f\n", 
		 (unsigned int)*dev, (double)*inode));

	return True;
}


/****************************************************************************
see if a oplock message is waiting
****************************************************************************/
static BOOL linux_oplock_msg_waiting(fd_set *fds)
{
	return signals_processed != signals_received;
}

/****************************************************************************
see if the kernel supports oplocks
****************************************************************************/
static BOOL linux_oplocks_available(void)
{
	int fd, ret;
	fd = open("/dev/null", O_RDONLY);
	if (fd == -1) return False; /* uggh! */
	ret = fcntl(fd, F_GETLEASE, 0);
	close(fd);
	return ret == F_UNLCK;
}


/****************************************************************************
setup kernel oplocks
****************************************************************************/
struct kernel_oplocks *linux_init_kernel_oplocks(void) 
{
	static struct kernel_oplocks koplocks;
        struct sigaction act;

	if (!linux_oplocks_available()) {
		DEBUG(3,("Linux kernel oplocks not available\n"));
		return NULL;
	}

        act.sa_handler = NULL;
        act.sa_sigaction = signal_handler;
        act.sa_flags = SA_SIGINFO;
        if (sigaction(RT_SIGNAL_LEASE, &act, NULL) != 0) {
		DEBUG(0,("Failed to setup RT_SIGNAL_LEASE handler\n"));
		return NULL;
        }

	koplocks.receive_message = linux_oplock_receive_message;
	koplocks.set_oplock = linux_set_kernel_oplock;
	koplocks.release_oplock = linux_release_kernel_oplock;
	koplocks.parse_message = linux_kernel_oplock_parse;
	koplocks.msg_waiting = linux_oplock_msg_waiting;
	koplocks.notification_fd = -1;

	DEBUG(3,("Linux kernel oplocks enabled\n"));

	return &koplocks;
}



#else
 void oplock_linux_dummy(void) {}
#endif /* HAVE_KERNEL_OPLOCKS_LINUX */